Download Systems biology of virus-host signaling network

BMB
reports
Invited Mini Review
Systems biology of virus-host signaling network interactions
Qiong Xue & Kathryn Miller-Jensen*
Department of Biomedical Engineering, Yale University, New Haven, CT 06511,USA
Viruses have evolved to manipulate the host cell machinery for
virus propagation, in part by interfering with the host cellular
signaling network. Molecular studies of individual pathways
have uncovered many viral host-protein targets; however, it is
difficult to predict how viral perturbations will affect the
signaling network as a whole. Systems biology approaches rely
on multivariate, context-dependent measurements and computational analysis to elucidate how viral infection alters host
cell signaling at a network level. Here we describe recent
advances in systems analyses of signaling networks in both
viral and non-viral biological contexts. These approaches have
the potential to uncover virus- mediated changes to host
signaling networks, suggest new therapeutic strategies, and
assess how cell-to-cell variability affects host responses to
infection. We argue that systems approaches will both improve
understanding of how individual virus-host protein interactions
fit into the progression of viral pathogenesis and help to
identify novel therapeutic targets. [BMB reports 2012; 45(4):
213-220]
INTRODUCTION
Cells respond to environmental changes by translating a complex network of protein interactions and biochemical signaling
reactions into functional responses (e.g., cytokine secretion or
cell death). In human diseases such as pathogenic viral infections and cancer, malfunctioning signaling networks cause
cells to incorrectly respond to stimuli and produce the diseased state (1, 2). Viruses and other pathogens often induce
malfunction by mimicking interaction domains of host proteins, which can rewire signaling networks and change cell responses for their own purposes (3). Frequently, viruses interact
with host proteins that have many interacting partners and/or
are central to many paths in the network (4), presumably because targeting these central interactions provides the most efficient means of changing host responses at a systems level.
*Corresponding author. Tel: 1-203-432-4265; Fax: 1-203-432-0030;
E-mail: [email protected]
http://dx.doi.org/10.5483/BMBRep.2012.45.4.213
Received 1 March 2012
Keyword: Signaling, Systems biology, Viral infection
http://bmbreports.org
Therefore, viral infection represents a valuable model for performing systems-level analyses of signaling network function;
and quantitative, systems-level approaches are needed to fully
understand the mechanisms of viral pathogenicity.
The advent of high throughput and multiplex techniques,
such as DNA microarrays (5), mass spectrometry (6), and highly multiparametric flow cytometry (7, 8), has enabled simultaneous experimental measurement of many components in a
biological system and contributed to the growing field of systems biology. Although initially applied primarily to gene regulation and genomic analysis, systems biology studies are increasingly applied to protein networks–including cell signaling
networks-due to the increased availability of high-throughput
techniques to probe large protein data sets (9). It is possible to
divide systems biology studies into two categories: static studies, which take a “snapshot” of a biological network under a
single condition, or limited set of conditions; and dynamic
studies, which measure time-dependent changes in the network following treatment with environmental stimuli or other
biological cues. Both approaches have the potential to significantly increase our understanding of the complex mechanisms involved in viral infection; however, dynamic systems
biology studies have been less widely pursued.
In the past few years, large-scale genomic and proteomic
methods have been used to uncover the complex network of
interactions that characterize viral infections of a host (Fig. 1).
For example, a physical, regulatory and functional network of
human-influenza H1N1 interactions constructed using a yeast
two-hybrid screen and a genome-wide analysis of gene expression, led to the identification of 1735 candidate genes that
might be involved in host response to influenza infection (10).
Similarly, three RNAi screens of the determinants of HIV infection implicated hundreds of host proteins that were not previously identified (11-13). And very recently, microarray analysis was used to compare altered gene expression regulation in
three different host cell systems following avian influenza
(H5N1) infection (14). Network analyses of these data sets provide a comprehensive and organizational picture of protein-protein or genetic interactions (usually undirected) that define the cellular state following viral infection. Such maps of
the virus-host interactome improve our understanding of the
global nature of viral infection on the host cell’s regulatory network; however they do not explain how virus–host interactions
function at a systems level to affect pathogenesis.
In contrast, dynamic systems biology approaches attempt to
BMB reports
213
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
describe the mechanisms of regulation between components
within the system by measuring cell responses to stimuli, usually as a function of time or dose (15). The resulting data is
generally context-dependent (e.g., specific to a particular cell
type or condition) and multivariate, and an array of mathematical and computational modeling approaches have been used
to extract underlying regulatory mechanisms (16). Measuring
time-dependent responses to stimuli for defined experimental
systems has been applied over the years to study the effect of
viral infection or viral proteins on single signaling proteins or
pathways, generating invaluable information on the mechanism of action in the network (Fig. 1). However it is often difficult to predict the effect of these interactions in the larger
network. Application of dynamic systems biology approaches
to virus-host signaling interactions, in which multiple signals in
the network are measured simultaneously over time, may provide a better understanding of how a virus hijacks the host protein signaling network and wires signaling in favor of virus survival and replication (Fig. 1).
In this review, we will describe recent advances in quantitative, perturbation-based systems biology approaches applied to
1) defining the organization and function of virus-mediated
changes to host signaling networks; 2) designing and evaluating anti-viral therapies; and 3) analysis of cell-to-cell variability
and how it affects host response to viral infection. We highlight studies from both the viral and non-viral literature to demonstrate the range of computational approaches used to analyze dynamic, multivariate data to extract novel biological
mechanisms. We suggest that application of these approaches
to a broader range of viral infections would significantly improve our understanding of viral pathogenesis, and may uncover novel targets for antiviral therapies to treat the multitude
of viral infections that continue to confound the medical
community.
SYSTEMS APPROACHES TO ANALYZE VIRAL-MEDIATED
CHANGES TO SIGNALING NETWORK ORGANIZATION
AND FUNCTION
Fig. 1. Dynamic systems biology approaches to study virus-host interactions (center) complement large-scale studies of networks of virus-host protein-protein interactions (left) and molecular level studies
of single pathways targeted by infection (right). In this schematic of
an infected host cell’s signaling network, host proteins (white circles)
or viral proteins (black circles) are depicted as undirected interactions
(black), activating interactions (green), or inhibitory interactions (red).
Computational models (see Table 1) can be developed to understand
how viral infection alters downstream signaling responses.
Viruses alter normal regulation of host signaling processes, in
part by interacting with host signaling proteins to “rewire” the
network (3, 17). Protein interaction networks generated from
large-scale, high-throughput studies have provided topological
information about the complexity of these interactions.
However, such networks do not provide information about
how viral infection alters host cell responses, both directly and
in response to extracellular stimuli. Experimental data linking
Table 1. Glossary
Bayesian network
A form of graphical modeling that calculates the most probable set of interactions between a set of variables
(e.g., proteins) based on experimental measurements of these variables
Boolean logic model
A discrete modeling technique which calculates a binary state of a protein (either "active" or "inactive")
based on its dependence on the states of other proteins in the network
Cluster analysis (clustering)
A type of model that classifies objects (e.g., biological sample or conditions) based on the similarity of measured characteristics (e.g., activation state of proteins over time)
Graphical Gaussian modeling
A graphical modeling technique that calculates the dependecy of each variable on all other variables in the
network
Hierarchical clustering
A subgroup of clustering analysis that arranges clusters in a hierarchical tree
Multiple linear regression
A model in which the dependent variable is a linear combination of one or more independent variables
Artificial neural network
A learning-based approach that infers a function relating an output (e.g., biological state) to input variables
(e.g., protein measurements) based on a set of input-output observations
Ordinary differential equation model A system of equations that defines the time-dependent change of each variable in the network in terms of the
other variables and a set of parameters that characterize the system
Partial least squares regression
A modeling technique for relating independent and dependent variables via regression when the number of
measured variables is greater than the number of experimental observations
Stochastic fluctuation
Random fluctuation due to small numbers of biological molecules (e.g., transcripts or proteins) that can lead
to non-genetic heterogeneity
Superposition of effect
A model that quantifies interactions between single agents (both synergy and antagonism)
214 BMB reports
http://bmbreports.org
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
extracellular cues with downstream signals and responses
(so-called “cue-signal-response” approaches) are extremely valuable for understanding which protein interactions are functionally linked to host cell responses. For example, 82 protein
interactions are implicated in the immediate-early response of
human cells to a range of cytokines based on literature information of the network (18). However, experimental data
from time courses of signals and responses gathered in liver
cells following stimulation with the same cytokine cues reveals
a different picture. To assess the functional interactions evident
from the cue-signal-response data, the data were fit to a
Boolean logic model, which essentially adds direction of influence to protein interaction diagrams (Table 1). The resulting
model had greater predictive power than the literature-derived
network, while reducing the total number of connections, suggesting that many interactions found in the literature were not
functional in the specific experimental system considered (i.e.,
liver hepatocarcinoma cells) (18). Similarly, experimental data
measuring time-dependent changes in the viral systems of interest will be necessary to extract the most useful information
from the viral–host protein interactome data rapidly becoming
available. Below we describe a range of systems approaches
that are well suited to address this challenge.
Several recent studies have directly measured time-dependent changes across multiple pathways in response to virus
infection. For example, in an effort to understand why SIV infection is pathogenic in Asian macaques (a non-natural host)
and not pathogenic in African green monkeys (a natural host),
Lederer et al. measured viral-induced changes in gene expression over time and in multiple tissues derived from each
host (19). The dynamic measurements showed that both infections induce a strong type I interferon response, but the interferon response peaks and falls in the non-pathogenic infection, while a sustained response is associated with
pathogenesis. Another non-viral study measured a time course
of global gene expression using microarrays data, and reported
that infection of wild type and mutant Salmonella induced significantly different patterns of host signaling within phosphatidylinositol, CCR3, Wnt, TGF-β and actin regulation pathways
(20). In contrast to gene expression, few studies have measured multivariate changes in signaling protein activity following viral infection. A recent pioneering investigation of host
cell signaling responses to coxsackievirus B3 (CVB3) infection
sampled phospho-protein dynamics in the presence of single
or pairwise combinations of small molecule inhibitors (21).
The signaling network reconstructed using graphical Gaussian
modeling (Table 1) uncovered an extracellular autocrine circuit involving TNF and IL-1 necessary for CVB3 cardiotoxicity,
and suggested a potential strategy for therapy (see below).
Another effective way to identify viral–host protein interactions that alter underlying signaling mechanisms is to measure virus-induced changes in signaling and phenotypic responses to extracellular stimuli. For example, in cells infected
with an attenuated adenoviral vector (Adv), Adv infection satuhttp://bmbreports.org
rated Akt pro-survival signaling and blocked insulin-mediated
anti-apoptotic signaling in cells treated with TNF, causing
apoptosis in host cells (2). In another study comparing HIV-1
infected and uninfected monocytes, HIV-1 infection significantly changed monocyte activation in response to granulocyte-macrophage colony-stimulating factor (GM-CSF) via
down-regulation of Jak/STAT signaling but enhancement of
MAPK signaling (22), resulting in defective antigen presentation. These studies focused on a single pathway, however
the extent of viral-mediated changes could be more fully understood by measuring changes across multiple pathways in
the network and using computational methods to analyze the
experimental data. For example, in an effort to understand system-level changes in a cell signaling network induced by cancer, secretion of 50 cytokines and measurements of 17 intracellular signals were compared between primary hepatocytes
and transformed liver cell lines following stimulation with 7 inducers of inflammation, innate immunity and proliferation in
the presence or absence of 7 small molecule inhibitors of specific pathways (23). The authors used multiple linear regression (Table 1) to correlate the strength of interaction between cytokine cues, signaling proteins and cytokine secretion
responses, and overlaid these onto literature-constructed signaling network diagrams. This study revealed significant differences in the engagement of toll-like receptors and NF-κB dependent cytokine and chemokine release in the normal and
transformed cells (23), demonstrating the underlying changes
in signaling that produce the cancer phenotype. Such an approach could be analogously applied to compare virally-infected and uninfected cells.
Another approach is to compare systems-level signaling responses between viruses and environmental stimuli that act on
the same network. For example, a time course analysis of genome-wide expression patterns induced in Jurkat cells in response to expression of HIV Nef or TCR stimulation with anti-CD3 demonstrated that Nef closely mimics T cell activation,
and that this may prime the T cell for HIV infection (24). In another study applied to a non-viral pathogen, Franke et al. compared how Helicobacter pylori (H. pylori) and hepatocyte
growth factor (HGF) induce activation of c-Met receptor signaling (25). The study, which applied a variation of Boolean modeling (Table 1), revealed protein targets downstream of c-Met
activated only during H. pylori infection, but not following
HGF stimulation in uninfected cells. Another recent non-viral
pathogen study similarly compared growth factor- and
Salmonella enterica-induced signaling events in host cells using a global, temporal phospho-proteomic mass spectrometry
data set and clustering analysis (Table 1) (26). The authors
demonstrated that changes in phosphorylation patterns in the
infected cells were a result of altered Akt, protein kinase C and
Pim activity, and many altered phosphorylation events depended on a single bacterial effector protein (26). Both studies
thus suggest possible intervention targets for diseases induced
by the respective bacterial infections, and have clear parallels
BMB reports
215
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
to viral infection.
Following an experimental systems analysis of how network
function is altered after viral infection, the information encoded in the resulting multivariate signaling measurements
can be used to build models that predict biological responses
in the presence and absence of infection. The first study to
demonstrate this approach in a non-viral system analyzed the
molecular basis of apoptosis induction in response to combinations of both pro- and anti-apoptotic stimuli (TNF, EGF and
insulin). Nineteen intracellular measurements and four distinct
apoptotic outputs were quantified (27), and a partial least
squares regression model (PLSR; Table 1) constructed from
these data identified the underlying combination of time-dependent signals associated with apoptosis. Furthermore, the
model predicted the effect of small molecule inhibition of autocrine feedback loops induced by TNF, results that were subsequently confirmed experimentally. This same type of model,
based on information included in multivariate measurements
of signaling responses, was able to correctly predict how infection with an adenoviral vector altered apoptotic responses
to TNFα in different cell types (28). Measurements of signaling
activation and apoptosis in colon carcinoma cells in response
to combinations of TNF and adenoviral vector infection or
IFN-γ were used to build a PLSR model that could accurately
predict the increases in TNF-mediated apoptosis observed in
infected cells. Surprisingly, the same model could also correctly predict how infection changed apoptotic responses in
other epithelial cell types (28). The finding that a model based
on multivariate signaling measurements in one cell type could
predict a wide range of virus-altered responses across divergent epithelial cell lines suggests a highly promising application to a broader range of viral infection studies. Dynamic
systems approaches that accurately predict cell-specific responses following viral infection could be used to test the outcomes of viral infection and anti-viral drug administration in
silico, before pursuing different approaches experimentally.
SYSTEMS APPROACHES TO DESIGN AND EVALUATE
ANTI-VIRAL THERAPIES
An improved understanding of how signal transduction pathways are altered by viral infection can be extended to reveal
how manipulation of these pathways will change cellular responses to ligands or drugs, and therefore inform novel strategies for therapy. In the following studies, dynamic systems biology approaches were used to reveal network regulatory
mechanisms which have significant clinical relevance.
Evaluation of drug efficacy and safety is a major endeavor in
the pharmaceutical industry. Enormous effort has been directed into understanding the mechanisms of anti-viral drugs,
such as anti-HCV drugs, interferon and ribavirin (29-31).
However, these studies focused on how drugs affected isolated
molecules or signaling pathways, and little was known about
the effect on signaling pathways in the context of the larger
216 BMB reports
network. In contrast, studies based on dynamic, multivariate
measurements of multiple pathways can provide a comprehensive evaluation of drug action. Mitsos et al. monitored
drug-induced signaling alterations at a systems level by measuring activation of thirteen key phospho-proteins in the presence or absence of EGFR inhibitors in a hepatocarcinoma cell
line (32). By integrating their data using a Boolean-based model, they confirmed the main targets of four drugs as well as uncovered several unknown off-target effects, demonstrating an
efficient way to determine drug selectivity and potency.
Similarly, a Boolean model of growth factor receptor activities
in response to different ligands or drugs discovered a poorly
documented off target effect of TPCA-1, an IκB inhibitor (33).
Similar approaches may demonstrate promising applications to
investigate anti-viral drug actions.
Current anti-viral drugs generally target viral factors, such as
the neuraminidase or the M2 ion channel of influenza (34) and
protease, integrase or reverse transcriptase in HIV-1 (35).
However, in part due to the emergence of drug-resistant
strains, increasingly pre-clinical approaches focus on host cellular factors or pathways that affect virus replication (36).
Given the complexity of virus–host interactions as described
above, systems-level studies are well suited for evaluating such
strategies. For example, recent research on novel anti-viral
therapies to treat Hepatitis C virus (HCV) have focused on host
sterol and protein prenylation pathways (37). To date, however, HCV therapies targeting the sterol pathway have not
yielded satisfactory results. To address this problem, Owens et
al. adopted a systems biology approach to study HCV replication in host cells treated with single or pairwise combinations of 16 chemical inhibitors that targeted the sterol and protein prenylation pathways (37). A superposition of effect model
(Table 1) revealed that the causes underlying failure of the
therapy were from complicated pathway regulation: sterol
pathway inhibition often results in host toxicity and epistatic
side effects. More interestingly, the model demonstrated that
high synergistic inhibition of HCV replication can be achieved
by combinatorial targeting of two downstream enzymes in the
sterol pathway, revealing a potential strategy for HCV therapy.
As demonstrated by the HCV study, combinatorial drug
therapy is an approach increasingly used to overcome viral
drug resistance by targeting multiple factors or pathways necessary for infection (38). For the treatment of many viruses, including HIV, HCV and Influenza, combinatorial drug administration has shown increased and broadened efficacy (38-40).
However, not all drugs can be combined in an effective way
without increased toxicity, and therefore investigations into the
combined effects of two or more anti-viral drugs are especially
significant. Dynamic systems approaches combined with signaling perturbation and computational modeling offer a
unique and efficient way to estimate combined drug actions.
For example, a strategy called pairwise agonist scanning
trained a neural network model (Table 1) based on experimental data sets of human platelet responses to all single
http://bmbreports.org
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
and pairwise combinations of input agonists (41). The model
was able to accurately predict responses from combinations of
three to six inputs, indicating the potential of such models to
provide useful in silico data about combinations of physiological inputs not yet tested experimentally.
Systems biology studies can also lead to unintuitive predictions of network regulatory mechanisms during virus infection and therefore guide therapeutic strategy. Oncolytic adenovirus has been tested in clinical trials for cancer therapy,
but efficacy has so far been inadequate. Inhibition of MEK has
the potential to enhance adenovirus infection but also decreases adenovirus replication, making it unclear how to optimize this strategy (42). To resolve this dilemma, Bagheri et al.
constructed an ordinary differential equation model (Table 1)
based on virus receptor expression, cell viability and proliferation, and virus infection and replication in order to test
various treatment strategies (42). Modeling dynamic cancer
cell activities in response to MEK inhibition and adenovirus infection predicted that optimal oncolytic treatments could be
achieved by simultaneous treatment with oncolytic adenovirus
infection and MEK inhibitors.
Another common problem leading to unexpected consequences of therapy, both in viral infection and other diseases
such as cancer, is the existence of autocrine loops, in which
secreted cytokines induce unexpected responses downstream
of an initial stimulus. In a study on TNF-induced human epithelial cell signaling (43), a classifier based regression model
built on ∼8,000 intracellular measurements revealed that
TNF, a pro-death stimulus, activated a sequential release of
three cytokines, both pro- and anti-apoptotic in function. The
signaling induced by the autocrine cascade specified the extent of apoptosis induced by TNF. This study provided insight
into why certain anti-tumor drugs targeting only one component of this response are sometimes ineffective (43). Autocrine
circuits similarly confound effective treatment strategies for
coxsackievirus B3 (CVB3) infection (see above). The same systems biology study that uncovered an extracellular autocrine
circuit involving TNF and IL-1 necessary for CVB3 cardiotoxicity, demonstrated that blocking this positive feedback circuit
significantly inhibited CVB3 replication and improved host
cell viability (21). These discoveries have important parallels
in host-viral systems as several viruses have been reported to
interact with TNF/TNFR pathways to favor virus replication
(Epstein-Barr virus) or trigger the killing of bystander cells
(Hepatitis B virus and HCV) (44). Thus, future investigations of
the role of extracellular autocrine circuits in host-viral systems
may provide invaluable insights on viral pathogenesis and the
development of efficient anti-viral therapy.
SYSTEMS APPROACHES TO ASSESS THE EFFECT OF
CELL-TO-CELL VARIABILITY IN VIRAL INFECTION
One question facing much of biology is why cells exposed to
the same stimuli or pathogen demonstrate different phenotypic
http://bmbreports.org
outcomes (45). Significant effort has been and continues to be
made in understanding how cells integrate and respond to environmental perturbations at a population level. However,
population-averaged measurements often fail to recognize important information that originated from population heterogeneity. In the past, cell-to-cell heterogeneity was often considered an obstacle to interpreting population responses induced
by extracellular stimuli. However, quantitative systems analyses of signaling networks at a single cell level are using the information encoded in cell population heterogeneity to yield
valuable insights into understanding signaling network function, including in response to viral infection.
Cell-to-cell heterogeneity is seen at all levels of the viral lifecycle, including 1) heterogeneity in the incidence of infection
itself (46); 2) variation in host cell response following infection
(47); and 3) heterogeneity in viral fate, such as replication versus latency (48, 49). To understand cell-to-cell variability in viral infection, Snijder et al. collected large numbers of single-cell measurements of each cells’ local environment (e.g.,
cell population size, local cell density, single cell position, cell
size, mitotic state and apoptotic state) from three types of cells
infected by one of three viruses (46). Using graphical Gaussian
modeling and linear regression (Table 1), the researchers built
a model that was able to predict heterogeneous infection patterns based on each cells’ population context. Moreover,
Bayesian network learning (Table 1) revealed a novel infection
mechanism for SV40 virus, which was further validated by
experiments. This study not only revealed how heterogeneity
of cellular activities during virus infection accounts for variability in infection, but also demonstrated how systems biology
approaches at the single cell level can improve understanding
of viral infection mechanisms.
Viral infection can also induce cell-to-cell viability in activation of signaling pathways. A recent study in mouse cells infected with Sendal virus demonstrated that cells infected with
similar levels of virus showed stochastic expression of interferon-β (IFNβ) and other virus-inducible genes (50). The reason
for the stochastic IFNβ response appeared to be due to
cell-to-cell variability in host protein factors mediating all levels of the infection process, including sensors of infection, signaling proteins, and transcription factors (50). Interestingly,
naturally-occurring variations in protein levels were also
shown to account for cell-to-cell heterogeneity in apoptotic responses to treatment with TRAIL (51). This suggests that resistance to treatment exhibited by many cancer cell populations
may not be genetically based, a result that may also apply to
anti-viral treatment. Such natural stochastic fluctuations (Table
1) may also account for observed differences in cell fate, such
as the replication-versus-latency decision of HIV-1 (48, 49).
Stochastic fluctuations in the viral protein Tat were shown to
result in different patterns of gene expression in populations of
Jurkat cells clonally infected with a minimal HIV vector (52,
53). Whether or not variability in host proteins also contributes
to this viral fate decision remains to be worked out.
BMB reports
217
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
Measurements of signaling proteins at the single cell level
increasingly reveal network functions and mechanisms that are
masked in the population-level measurements. For example,
uniform stimulation of a PC12 cell population by nerve growth
factor (NGF) causes some cells to differentiate and some cells
to proliferate. A two-dimensional phospho-ERK-phospho-Akt
single cell response map constructed by single cell imaging
analysis of cells treated with NGF revealed a distinct boundary
separating differentiating and proliferating cells (54). The same
pERK–pAkt boundary separated proliferation-versus-differentiation fate decision in response to other growth factors,
which suggested a mechanistic strategy by which cells regulate
differentiation within the population (54). Interestingly, another recent study exploited the cell-to-cell variability of viral infection to quantify signaling responses to heterogeneous levels
of MYC protein (55). By using single cell imaging to quantify
signal activation in response to a wide range of overexpressed
MYC delivered by an adenoviral vector, the authors revealed a
biphasic pattern in the activation of MYC’s downstream effector, E2f1, accounting for conflicting reports of the effect of
MYC overexpression.
Increasingly, single cell studies at a systems level are revealing the complexity of signaling responses in cells of the immune system. Given the close relationship between viral infection and the resulting host immune response, these studies
may provide important insights to improving anti-viral therapies and vaccine development. For example, a serial, time-dependent, single cell analysis of IFN-gamma, IL-2 and TNF-alpha secretion by primary human T cells in response to activating stimuli showed that cytokine release is asynchronous and
that the majority of cells secrete only one cytokine at a time
(56). Furthermore, the study showed that T cells release cytokines following a programmatic pattern which is associated
with the differentiated state of the cell. A similar study applied
information theory to investigate fluctuations in protein secretion from single human macrophage cells in response to LPS.
By characterizing the fluctuations of different secreted proteins
in a single cell and in small cell colonies, the authors were
able to construct a protein-protein interaction network and
quantitatively predict the role of perturbations (57). Notably,
both of these studies were conducted in microfluidic devices,
which increasingly facilitate quantification of signaling and cytokine responses in single cells due to the sensitivity provided
by miniaturization of assays (58).
Single cells from the same viral-infected population exhibit
differences in the number of produced virus particles spanning
over 300-fold (59), and this diversity will surely impact infection dynamics and treatment strategies. Therefore, studies of
virus–host interactions will be especially enhanced by the
growing availability of multiplexed experimental measurements of proteins in single cells, and computational approaches to derive biological understanding from this multitude of data.
218 BMB reports
CONCLUSIONS
Viral infection is a systems-level perturbation, and therefore
systems biology approaches are naturally suited to studying
the complexity of viral pathogenesis. We have highlighted a
number of innovative examples from both the viral and non-viral literature, which demonstrate the potential for dynamic systems biology signaling studies–characterized by time-variant,
context-dependent responses to viral infection–to complement
both molecular, single pathway efforts and high-throughput,
large-scale analyses of viral–host interactomes (Fig. 1). We expect that systems biology approaches designed to define signaling network organization, discover novel regulatory mechanisms, and predict phenotypic responses to viral infection will
uncover many novel mechanisms underlying viral pathogenesis and point to promising strategies for anti-viral therapy.
REFERENCES
1. Irish, J. M., Hovland, R., Krutzik, P. O., Perez, O. D.,
Bruserud, O., Gjertsen, B. T. and Nolan, G. P. (2004)
Single cell profiling of potentiated phospho-protein networks in cancer cells. Cell 118, 217-228.
2. Miller-Jensen, K., Janes, K. A., Wong, Y. L., Griffith, L. G.
and Lauffenburger, D. A. (2006) Adenoviral vector saturates Akt pro-survival signaling and blocks insulin-mediated rescue of tumor necrosis-factor-induced
apoptosis. J. Cell Sci. 119, 3788-3798.
3. Pawson, T. and Warner, N. (2007) Oncogenic re-wiring of
cellular signaling pathways. Oncogene 26, 1268-1275.
4. Dyer, M. D., Murali, T. M. and Sobral, B. W. (2008) The
landscape of human proteins interacting with viruses and
other pathogens. PLoS Pathog. 4, e32.
5. Kingsmore, S. F. (2006) Multiplexed protein measurement: technologies and applications of protein and antibody arrays. Nat. Rev. Drug Discov. 5, 310-320.
6. Domon, B. and Aebersold, R. (2006) Mass spectrometry
and protein analysis. Science 312, 212-217.
7. Perfetto, S. P., Chattopadhyay, P. K. and Roederer, M.
(2004) Seventeen-colour flow cytometry: unravelling the
immune system. Nat. Rev. Immunol. 4, 648-655.
8. Ornatsky, O., Bandura, D., Baranov, V., Nitz, M., Winnik,
M. A. and Tanner, S. (2010) Highly multiparametric analysis by mass cytometry. J. Immunol. Methods 361, 1-20.
9. Cox, M. G., van der Zwaag, P. A., van der Werf, C., van
der Smagt, J. J., Noorman, M., Bhuiyan, Z. A., Wiesfeld,
A. C., Volders, P. G., van Langen, I. M., Atsma, D. E.,
Dooijes, D., van den Wijngaard, A., Houweling, A. C.,
Jongbloed, J. D., Jordaens, L., Cramer, M. J., Doevendans,
P. A., de Bakker, J. M., Wilde, A. A., van Tintelen, J. P.
and Hauer, R. N. (2011) Arrhythmogenic right ventricular
dysplasia/cardiomyopathy: pathogenic desmosome mutations in index-patients predict outcome of family screening: Dutch arrhythmogenic right ventricular dysplasia/cardiomyopathy genotype-phenotype follow-up study. Circulation 123, 2690-2700.
10. Shapira, S. D., Gat-Viks, I., Shum, B. O., Dricot, A., de
http://bmbreports.org
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Grace, M. M., Wu, L., Gupta, P. B., Hao, T., Silver, S. J.,
Root, D. E., Hill, D. E., Regev, A. and Hacohen, N. (2009)
A physical and regulatory map of host-influenza interactions reveals pathways in H1N1 infection. Cell 139,
1255-1267.
Brass, A. L., Dykxhoorn, D. M., Benita, Y., Yan, N.,
Engelman, A., Xavier, R. J., Lieberman, J. and Elledge, S. J.
(2008) Identification of host proteins required for HIV infection through a functional genomic screen. Science 319,
921-926.
Konig, R., Zhou, Y., Elleder, D., Diamond, T. L., Bonamy,
G. M., Irelan, J. T., Chiang, C. Y., Tu, B. P., De Jesus, P.
D., Lilley, C. E., Seidel, S., Opaluch, A. M., Caldwell, J. S.,
Weitzman, M. D., Kuhen, K. L., Bandyopadhyay, S.,
Ideker, T., Orth, A. P., Miraglia, L. J., Bushman, F. D.,
Young, J. A. and Chanda, S. K. (2008) Global analysis of
host-pathogen interactions that regulate early-stage HIV-1
replication. Cell 135, 49-60.
Zhou, H., Xu, M., Huang, Q., Gates, A. T., Zhang, X. D.,
Castle, J. C., Stec, E., Ferrer, M., Strulovici, B., Hazuda, D.
J. and Espeseth, A. S. (2008) Genome-scale RNAi screen
for host factors required for HIV replication. Cell Host.
Microbe. 4, 495-504.
McDermott, J. E., Shankaran, H., Eisfeld, A. J., Belisle, S.
E., Neuman, G., Li, C., McWeeney, S., Sabourin, C.,
Kawaoka, Y., Katze, M. G. and Waters, K. M. (2011)
Conserved host response to highly pathogenic avian influenza virus infection in human cell culture, mouse and
macaque model systems. BMC Syst. Biol. 5, 190.
Albeck, J. G., MacBeath, G., White, F. M., Sorger, P. K.,
Lauffenburger, D. A. and Gaudet, S. (2006) Collecting and
organizing systematic sets of protein data. Nat. Rev. Mol.
Cell Biol. 7, 803-812.
Janes, K. A. and Lauffenburger, D. A. (2006) A biological
approach to computational models of proteomic networks. Curr. Opin. Chem. Biol. 10, 73-80.
Davey, N. E., Trave, G. and Gibson, T. J. (2011) How viruses hijack cell regulation. Trends. Biochem. Sci. 36,
159-169.
Saez-Rodriguez, J., Alexopoulos, L. G., Epperlein, J.,
Samaga, R., Lauffenburger, D. A., Klamt, S. and Sorger, P.
K. (2009) Discrete logic modelling as a means to link protein signalling networks with functional analysis of mammalian signal transduction. Mol. Syst. Biol. 5, 331.
Lederer, S., Favre, D., Walters, K. A., Proll, S., Kanwar, B.,
Kasakow, Z., Baskin, C. R., Palermo, R., McCune, J. M.
and Katze, M. G. (2009) Transcriptional profiling in pathogenic and non-pathogenic SIV infections reveals significant distinctions in kinetics and tissue compartmentalization. PLoS Pathog. 5, e1000296.
Lawhon, S. D., Khare, S., Rossetti, C. A., Everts, R. E.,
Galindo, C. L., Luciano, S. A., Figueiredo, J. F., Nunes, J.
E., Gull, T., Davidson, G. S., Drake, K. L., Garner, H. R.,
Lewin, H. A., Baumler, A. J. and Adams, L. G. (2011) Role
of SPI-1 secreted effectors in acute bovine response to
Salmonella enterica Serovar Typhimurium: a systems biology analysis approach. PLoS One 6, e26869.
Garmaroudi, F. S., Marchant, D., Si, X., Khalili, A.,
Bashashati, A., Wong, B. W., Tabet, A., Ng, R. T.,
Murphy, K., Luo, H., Janes, K. A. and McManus, B. M.
http://bmbreports.org
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
(2010) Pairwise network mechanisms in the host signaling
response to coxsackievirus B3 infection. Proc. Natl. Acad.
Sci. USA 107, 17053-17058.
Lee, A. W., Sharp, E. R., O'Mahony, A., Rosenberg, M. G.,
Israelski, D. M., Nolan, G. P. and Nixon, D. F. (2008)
Single-cell, phosphoepitope-specific analysis demonstrates cell
type- and pathway-specific dysregulation of Jak/STAT and
MAPK signaling associated with in vivo human immunodeficiency virus type 1 infection. J. Virol. 82, 3702-3712.
Alexopoulos, L. G., Saez-Rodriguez, J., Cosgrove, B. D.,
Lauffenburger, D. A. and Sorger, P. K. (2010) Networks inferred from biochemical data reveal profound differences
in toll-like receptor and inflammatory signaling between
normal and transformed hepatocytes. Mol. Cell Proteomics
9, 1849-1865.
Simmons, A., Aluvihare, V. and McMichael, A. (2001) Nef
triggers a transcriptional program in T cells imitating single-signal T cell activation and inducing HIV virulence
mediators. Immunity 14, 763-777.
Franke, R., Muller, M., Wundrack, N., Gilles, E. D.,
Klamt, S., Kahne, T. and Naumann, M. (2008) Host-pathogen systems biology: logical modelling of hepatocyte
growth factor and Helicobacter pylori induced c-Met signal transduction. BMC Syst. Biol. 2, 4.
Rogers, L. D., Brown, N. F., Fang, Y., Pelech, S. and
Foster, L. J. (2011) Phosphoproteomic analysis of
Salmonella-infected cells identifies key kinase regulators
and SopB-dependent host phosphorylation events. Sci.
Signal. 4, rs9.
Janes, K. A., Albeck, J. G., Gaudet, S., Sorger, P. K.,
Lauffenburger, D. A. and Yaffe, M. B. (2005) A systems
model of signaling identifies a molecular basis set for cytokine-induced apoptosis. Science 310, 1646-1653.
Miller-Jensen, K., Janes, K. A., Brugge, J. S. and Lauffenburger, D. A. (2007) Common effector processing mediates cell-specific responses to stimuli. Nature 448,
604-608.
Fahey, B., Hickey, B., Kelleher, D., O'Dwyer, A. M. and
O'Mara, S. M. (2007) The widely-used anti-viral drug interferon-alpha induces depressive- and anxiogenic-like effects in healthy rats. Behav. Brain Res. 182, 80-87.
Guido, M., Bortolotti, F., Jara, P., Giacomelli, L., Fassan,
M., Hierro, L., Nebbia, G., Zancan, L. and Rugge, M.
(2006) Liver steatosis in children with chronic hepatitis C.
Am. J. Gastroenterol. 101, 2611-2615.
Bortoletto, G., Scribano, L., Realdon, S., Marcolongo, M.,
Mirandola, S., Franceschini, L., Bonisegna, S., Noventa, F.,
Plebani, M., Martines, D. and Alberti, A. (2010)
Hyperinsulinaemia reduces the 24-h virological response
to PEG-interferon therapy in patients with chronic hepatitis
C and insulin resistance. J. Viral. Hepat. 17, 475-480.
Mitsos, A., Melas, I. N., Siminelakis, P., Chairakaki, A. D.,
Saez-Rodriguez, J. and Alexopoulos, L. G. (2009) Identifying drug effects via pathway alterations using an integer linear programming optimization formulation on phosphoproteomic data. PLoS Comput. Biol. 5, e1000591.
Saez-Rodriguez, J., Alexopoulos, L. G., Zhang, M., Morris,
M. K., Lauffenburger, D. A. and Sorger, P. K. (2011)
Comparing signaling networks between normal and transformed hepatocytes using discrete logical models. Cancer
BMB reports
219
Systems biology of virus-host signaling network interactions
Qiong Xue and Kathryn Miller-Jensen
Res. 71, 5400-5411.
34. Ludwig, S. (2011) Disruption of virus-host cell interactions
and cell signaling pathways as an anti-viral approach
against influenza virus infections. Biol. Chem. 392,
837-847.
35. Kirchmair, J., Distinto, S., Liedl, K. R., Markt, P., Rollinger,
J. M., Schuster, D., Spitzer, G. M. and Wolber, G. (2011)
Development of anti-viral agents using molecular modeling and virtual screening techniques. Infect. Disord. Drug
Targets. 11, 64-93.
36. Peng, X., Chan, E. Y., Li, Y., Diamond, D. L., Korth, M. J.
and Katze, M. G. (2009) Virus-host interactions: from systems biology to translational research. Curr. Opin.
Microbiol. 12, 432-438.
37. Owens, C. M., Mawhinney, C., Grenier, J. M., Altmeyer,
R., Lee, M. S., Borisy, A. A., Lehar, J. and Johansen, L. M.
(2010) Chemical combinations elucidate pathway interactions and regulation relevant to Hepatitis C replication.
Mol. Syst. Biol. 6, 375.
38. Pirrone, V., Thakkar, N., Jacobson, J. M., Wigdahl, B. and
Krebs, F. C. (2011) Combinatorial approaches to the prevention and treatment of HIV-1 infection. Antimicrob.
Agents. Chemother. 55, 1831-1842.
39. Arase, Y. (2011) Viral hepatitis: Impact of adherence to combination therapy for hepatitis C. Nat. Rev. Gastroenterol.
Hepatol. 8, 248-249.
40. Tsiodras, S., Mooney, J. D. and Hatzakis, A. (2007) Role
of combination antiviral therapy in pandemic influenza
and stockpiling implications. BMJ 334, 293-294.
41. Chatterjee, M. S., Purvis, J. E., Brass, L. F. and Diamond, S.
L. (2010) Pairwise agonist scanning predicts cellular signaling responses to combinatorial stimuli. Nat. Biotechnol.
28, 727-732.
42. Bagheri, N., Shiina, M., Lauffenburger, D. A. and Korn, W.
M. (2011) A dynamical systems model for combinatorial
cancer therapy enhances oncolytic adenovirus efficacy by
MEK-inhibition. PLoS Comput. Biol. 7, e1001085.
43. Janes, K. A., Gaudet, S., Albeck, J. G., Nielsen, U. B.,
Lauffenburger, D. A. and Sorger, P. K. (2006) The response of human epithelial cells to TNF involves an inducible autocrine cascade. Cell 124, 1225-1239.
44. Herbein, G. and O'Brien, W. A. (2000) Tumor necrosis
factor (TNF)-alpha and TNF receptors in viral pathogenesis. Proc. Soc. Exp. Biol. Med. 223, 241-257.
45. Niepel, M., Spencer, S. L. and Sorger, P. K. (2009)
Non-genetic cell-to-cell variability and the consequences
for pharmacology. Curr. Opin. Chem. Biol. 13, 556-561.
46. Snijder, B., Sacher, R., Ramo, P., Damm, E. M., Liberali,
P. and Pelkmans, L. (2009) Population context determines
cell-to-cell variability in endocytosis and virus infection.
220 BMB reports
Nature 461, 520-523.
47. Zawatzky, R., De Maeyer, E. and De Maeyer-Guignard, J.
(1985) Identification of individual interferon-producing
cells by in situ hybridization. Proc. Natl. Acad. Sci. U.S.A
82, 1136-1140.
48. Singh, A. and Weinberger, L. S. (2009) Stochastic gene expression as a molecular switch for viral latency. Curr.
Opin. Microbiol. 12, 460-466.
49. Miller-Jensen, K., Dey, S. S., Schaffer, D. V. and Arkin, A.
P. (2011) Varying virulence: epigenetic control of expression noise and disease processes. Trends. Biotechnol.
29, 517-525.
50. Zhao, M., Zhang, J., Phatnani, H., Scheu, S. and Maniatis,
T. (2012) Stochastic Expression of the Interferon-beta
Gene. PLoS Biol. 10, e1001249.
51. Spencer, S. L., Gaudet, S., Albeck, J. G., Burke, J. M. and
Sorger, P. K. (2009) Non-genetic origins of cell-to-cell variability in TRAIL-induced apoptosis. Nature 459, 428-432.
52. Weinberger, L. S., Burnett, J. C., Toettcher, J. E., Arkin, A.
P. and Schaffer, D. V. (2005) Stochastic gene expression
in a lentiviral positive-feedback loop: HIV-1 Tat fluctuations drive phenotypic diversity. Cell 122, 169-182.
53. Burnett, J. C., Miller-Jensen, K., Shah, P. S., Arkin, A. P.
and Schaffer, D. V. (2009) Control of stochastic gene expression by host factors at the HIV promoter. PLoS Pathog.
5, e1000260.
54. Chen, J. Y., Lin, J. R., Cimprich, K. A. and Meyer, T.
(2012) A Two-Dimensional ERK-AKT Signaling Code for
an NGF-Triggered Cell-Fate Decision. Mol. Cell. 45,
196-209.
55. Wong, J. V., Yao, G., Nevins, J. R. and You, L. (2011)
Viral-mediated noisy gene expression reveals biphasic
E2f1 response to MYC. Mol. Cell 41, 275-285.
56. Han, Q., Bagheri, N., Bradshaw, E. M., Hafler, D. A.,
Lauffenburger, D. A. and Love, J. C. (2012) Polyfunctional
responses by human T cells result from sequential release
of cytokines. Proc. Natl. Acad. Sci. U.S.A. 109, 1607-1612.
57. Shin, Y. S., Remacle, F., Fan, R., Hwang, K., Wei, W.,
Ahmad, H., Levine, R. D. and Heath, J. R. (2011) Protein
signaling networks from single cell fluctuations and information theory profiling. Biophys. J. 100, 2378-2386.
58. Ma, C., Fan, R., Ahmad, H., Shi, Q., Comin-Anduix, B.,
Chodon, T., Koya, R. C., Liu, C. C., Kwong, G. A., Radu,
C. G., Ribas, A. and Heath, J. R. (2011) A clinical microchip for evaluation of single immune cells reveals high
functional heterogeneity in phenotypically similar T cells.
Nat. Med. 17, 738-743.
59. Timm, A. and Yin, J. (2012) Kinetics of virus production
from single cells. Virology 424, 11-17.
http://bmbreports.org